We conducted an intensive fish survey in the tailwater reach of a large Ozark river 30 years after its impoundment and compared the recent fish assemblage with those prior to impoundment and shortly (4 years) after impoundment. Our primary objective was to assess whether relatively short‐term monitoring following dam construction can adequately quantify the long‐term effects of impoundment on downstream riverine fishes. The preimpoundment survey (1962–1963) described a fish assemblage composed of warmwater fish species, predominantly Cyprinidae, Ictaluridae, Centrarchidae, and Percidae. Yoke darter Etheostoma juliae (34%), central stoneroller Campostoma anomalum (24%), and Ozark madtom Noturus albater (7%) were the most abundant species. The postimpoundment surveys of 1965–1966 and 1968 documented immediate changes in the fish assemblage. No Ozark madtoms and only four yoke darters were collected shortly after impoundment. Central stonerollers accounted for 45–50% of the fish collected, and both short‐term postimpoundment surveys collected five species of darters (Percidae) that accounted for 41–42% of the fish collected. Thirty years after impoundment, we found that the tailwater fish assemblage was composed almost entirely of coldwater species. Ozark sculpin Cottus hypselurus and four species of introduced trout (Salmonidae) accounted for 98% of the fish assemblage by number during the 1995–1997 surveys. The rank abundance of species was negatively correlated between our survey and the preimpoundment survey but not between our survey and the short‐term postimpoundment surveys. Many species that we collected (54%) are habitat generalists, and we did not collect 77% of the fluvial‐specialist species that were present in historical collections. All postimpoundment surveys documented dramatically reduced species richness and diversity. We conclude that short‐term monitoring following impoundment is inadequate to determine the impact of dams on lotic fish assemblages and suggest long‐term postimpoundment monitoring to determine when a fish assemblage has stabilized.

We evaluated instream and riparian habitat rehabilitation that was completed following catastrophic flooding in the White River, below Beaver Dam, Arkansas. Most rehabilitation structures were designed to stabilize the river banks and increase cover for trout (Salmonidae) during high flows associated with hydroelectric power generation. We quantified trout response to rehabilitation at two spatial scales—microhabitat and river reach. At the microhabitat scale, brown trout Salmo trutta and rainbow trout Oncorhynchus mykiss occupied the deepest habitats available and were randomly associated with cover at low flow (about 1 m3/s). Principal‐component scores describing physical characteristics of brown trout and rainbow trout microhabitats were significantly different from available‐habitat scores at high flow (about 215 m3/s), when trout were strongly associated with velocity refugia near the river margins—habitats similar to those created by rehabilitation structures. At the reach scale, trout population size and structure were estimated in modified (700‐m) and reference (800‐m) reaches before and after rehabilitation. Total trout density and biomass in the modified reach increased after rehabilitation relative to that of the reference reach, evidence that the modified reach supported more fish after rehabilitation. Analyses stratified by salmonid species and size indicated that the observed effect was primarily due to rainbow trout and small trout (10.0–19.9 cm) of all species shifting their distributions into the modified reach. Our results suggest that instream and riparian habitat rehabilitation structures commonly applied to small streams are a valid management technique for large tailwater rivers. However, implementation in each system should be carefully evaluated, and management expectations for large trout should be conservative. Because of the observed benefit for small trout, placement of rehabilitation structures near spawning areas should be considered when management for wild trout is a priority. Finally, we suggest integration of instream and riparian habitat rehabilitation into broader management plans when applied to regulated rivers.

Bowfishing is an understudied method of fishing that appears to be legal throughout the United States. Therefore, species composition and harvest rates were determined at six bowfishing tournaments held in Arkansas at the lower White River, the Arkansas River at Lake Dardanelle, the Arkansas River at Piney Bay, Lake DeGray, Bull Shoals Lake, and Lake Ouachita between July 1999 and May 2000. A total of 3,280 fish were harvested at the six tournaments; of this total, 2,751 fish representing 19 species were identified. Total harvest per tournament ranged from 179 to 1,674 fish and from 6 to 12 species. Mean (±SD) harvest rate for tournament participants was 3.8 ± 1.1 fish/h; among tournament winners, the harvest rate was 7.7 ± 2.8 fish/h, which appears high compared with other sport fisheries (range = 0.28–2.59 fish/h). Five species accounted for 84% of fish harvested: spotted gar Lepisosteus oculatus, common carp Cyprinus carpio, shortnose gar L. platostomus, spotted sucker Minytrema melanops, and smallmouth buffalo Ictiobus bubalus. Rank number of each species harvested at the tournaments was correlated (P < 0.05) for only 4 of 15 pairwise comparisons, which suggests that harvest often varies by tournament. Tournaments held at the Arkansas and White rivers had correlated harvest, as did spring tournaments held at the Ouachita River drainage reservoirs (i.e., Lake DeGray and Lake Ouachita). Harvest of fish smaller than published size‐at‐maturity estimates was generally not problematic but appeared to be of greatest concern for smaller‐bodied catostomid species. This study indicates that tournament bowfishers have higher harvest rates than traditional rod‐and‐reel anglers. Results of this survey should provide baseline information that may assist natural resource agencies with management of bowfishing.